 |
 |
Testimonials
Control Limits vs Specification Limits
Many people do not understand the difference between process control limits and process specification limits. Using either in the wrong way can result in disastrous results, with poor customer service and financial loss to the organisation, so a good understanding of both is essential.
Process specification limits are usually derived from customer requirements, and should link to what the customer really wants. Process Control limits are derived from actual process performance, and are designed to show whether or not the process is in statistical control, they do not have any relationship to spec limits.
Process capability compares process performance with customer requirements (specification), and is expressed in the form of an index which shows how well the process is performing compared to the requirements. To avoid losses and customer dissatisfaction, specifications should be realistic and achievable, processes should be in statistical control and have a good capability value. We will explore these key concepts in this article.
Specification Limits
In the vast majority of cases, it will be the customer of the process that determines what the output of the process should look like. This could be in the form of:
A. A Dimensional Tolerance
This could be a 2 sided tolerance (a plus or minus allowance either side of a "nominal" or "target" value). For example, the head diameter of a screw head could have a target of 10mm with a plus and minus tolerance of 0.5mm. This allows the finished head diameter to fall anywhere between 9.5mm and 10.5 mm.
It could also be a one-sided tolerance, which gives a single allowance of movement away from a nominal or target value. For example, your boss would set a one-sided tolerance for your arrival time at work.
The nominal or target value would be set at 9.00am, and the tolerance applied to that would always be before 9.00am.
B. A Customer Service Tolerance
Almost on a daily basis, we all experience service delivery in one form or another. It could be the collection of refuge when you leave the house to go to work, it could be the bus or train journey to work or even the service you receive when you pop into Starbuck's on your way to work. We, as consumers, set a target for the level of service we deem to be acceptable and the limits within which we expect that service to fall. This is setting a Customer Specification. This specification is our view of what that service should look like when we receive it.
However, when dealing with business customers, a "wooly" specification will lead to conflict when issues arise as the expectation of the level of service expected is not the same on both sides. The best way of removing any level of ambiguity around specifications is to use strong, well targeted, Voice of the Customer, to understand EXACTLY what it is customers expect from you as either an individual or organization.
Capturing the Voice of the Customer (VOC) is a five stage process:
- Identify the information needed from your customers - this will form the basis of the questions you will ask your customers.
- Develop a research plan - plan how you are going to collect information from customers
- Collect Voice of the Customer (VOC) - Implement the plan developed in the previous step
- Analyse findings / define Critical Customer Requirements (CCR's) - review the results of the data collection exercise and draw conclusions. Determine what the customer requirements are.
- Develop process measures to monitor performance - Decide how you are going to identify whether or not the process output is going to meet the customer requirements.
The Critical Customer Requirements (CCR's) should be written in the form of a specification. To return back to our Starbucks example, the customer service expectation should be documented in a way that the employee at Starbucks knows exactly what they need to do to deliver the service required, and the customer knows exactly when the employee has delivered on the service expectation.
Process Control Limits
The first point to make about "Control Limits" is that they are in no way connected to the Customer Specification Limits. Process control limits are generated from the actual performance of the process.
All process output will have variation. This output variation is caused by changes in inputs. For example, time taken on your journey to work will be affected by factors such as weather, traffic, start time, speed and so on. Some small variation in journey time is inevitable. Sometimes there may be an accident on the route which causes a much larger variation, but this will be (hopefully!) unusual.
These different types of variation are called common and special cause variation. Common cause variation is variation that is normally present in processes and is to be expected. Special cause variation is not normally present and is unexpected. Process control limits calculated to highlight when special cause variation is present in a process, if the process output remains within the control limits then the process can be considered in control, if it goes outside control limits, then it is considered out of control and something unusual is happening.
The individual data points collected during the process are used to calculate 3 key data indicators:
- Process Average or "Mean" - This is the average value of all data points
- UCL (Upper Control Limit) - If only common cause variation is present, this line or limit on the chart represents a one in one thousand probability that a data point will fall above this limit. The UCL is set at +3 standard deviations from the mean.
- LCL (Lower Control Limit) - again, if only common variation is present, this line or limit on the chart represents a one in one thousand probability that a data point will fall below this limit. The LCL limit is set at -3 standard deviations from the mean.
A key observation to make here is that if a process has a high degree of variability, then the process control limits could actually be wider than the customer specification limits.
Process Capability
To determine whether a process has the ability to meet the customer specification, we can use a measure called "Process Capability". This is where both the process performance data and the customer specification limits are used together to calculate a process capability index.
Process Capability measures the output of a process and compares it to the customer specification limits, or can be considered as comparing Voice of the Customer with Voice of the Process. A good process will be centered on the nominal value, and the spread is low in relation to the specification limits.
DOWNLOAD : Process capability Capable Process.ppt
Poor capability is caused by either the spread of the process exceeding the specification limits, even if the process is centered; or by a process which may have a very small spread, but where the mean is not centered and is close to either the upper or lower spec limit resulting in process output being outside spec.
DOWNLOAD : Process capability Uncapable process.ppt
There are 4 common indices that relate to process capability. Cp, CpK, Pp, PpK.
Cp and Pp only compare process variation with spec limits and ignore centering, while Pp and PpK consider variation and also centering of the process. A full explanation of these indices is beyond the scope of this article, but Cp can be calculated using the formula below:
USL = Upper Spec Limit
LSL = Lower Spec Limit
S = standard deviation.
If Cp = 1 then the process will meet the specification 99.7% of the time, if it =1.33 then it will meet specification 99.99% of the time. Sometimes customers will define the Cp vale they require.
The trick to minimizing process loss is to make sure that any specification truly represents what the customer wants. Then, control the process so that it only exhibits common variation and is in control. Compare the process variation with the specification by calculating process capability and look for a value of 1.33 or better. If the index is below this, then work to improve the process.
Focusing on delivering a process that is centered on a target value and with a very low process spread will ensure that the probability of any product or service exceeding the specification limits is extremely low, and minimize financial loss.
The SigmaPro Green and Black Belt Lean Six Sigma courses give a thorough explanation of process variation, control and using capability to determine sigma performance. For more details of the upcoming courses see www.sigmapro.co.uk
Author Biography
Paul Martin
Quick Enquiry or Call Back
Get in touch, ask us anything! How can we help?
Six Sigma Yellow Belt | Six Sigma Green Belt | Six Sigma Black Belt | Six Sigma Master Black Belt
Locations : Lean Six Sigma Training in Birmingham | London | Newcastle | Leeds